Artificial Intelligence as an Inventor? A Legal Analysis of Thaler v Comptroller-General of Patents

Introduction

In the vast expansion of human innovation, inventions has always been the province of the mind. A spark of genius, a stroke of insight, a moment where the impossible bends to the will of human intellect. But what happens when the mind is no longer human? When creation itself is not the product of a human’s thought, but the product of a code. An intelligence that neither breathes nor dreams, yet conceives? This issue is at the heart of Thaler v. Comptroller-General of Patents, Designs and Trade Marks, where the fundamental principle that being an inventor is a fundamentally human activity, was contested.

The rapid advancement of artificial intelligence (AI) technology that we have seen in the past couple of years, raises significant ethical and legal concerns, especially in intellectual property. One of the most significant issues is whether AI systems may qualify as inventors under patent law. What raised awareness of this problem, is the landmark case of Thaler v. ComptrollerGeneral of Patents, that pushed the boundaries of common law states’ preexisting legal frameworks.

The argument revolves around the topic of whether and under what circumstances AI-generated ideas should be granted patent protection. The premise that natural persons are the only ones capable of inventing is the foundation of traditional patent rules all around the world, but as AI systems advance in sophistication, this presumption is coming under pressure. This problem has been debated by courts everywhere, with differing results in various countries.

The UK Supreme Court’s (UKSC) ruling in this case highlights the increasing conflict between legal frameworks and technical improvements while also reinforcing traditional interpretations of patent law. This analysis will critically examine the main legal contentions, the court’s reasoning, and the case’s greater implications for the development of artificial intelligence and intellectual property law.

Facts of the Case

An AI researcher named Dr. Stephen Thaler, filed an AI system DABUS (Device for the Autonomous Bootstrapping of Unified Sentience), that he owned, as the sole inventor in patent applications in several jurisdictions, including the UK, in 2018. Without human input, DABUS managed to produce two inventions on its own, which is an emergency beacon and a new kind of food container. Instead of claiming to be the inventor, Dr. Thaler contended that he was entitled to the rights of the patent, being the owner of DABUS.

However, the UK Intellectual Property Office (UKIPO) denied the application, claiming that inventors can only be natural persons in accordance with the Patents Act 1977. DABUS is unable to be considered an inventor, as it is an AI system without any legal personhood. Thaler’s patent applications were declared legally defective by the UKIPO and had to be withdrawn, as a result.

The High Court and the Court of Appeal maintained UKIPO’s verdict despite Dr. Thaler’s appeals. The case eventually made its way to the UK Supreme Court, where the main question was whether AI systems could qualify as an inventor under existing UK patent law.

Since Thaler sought comparable applications in the US, Europe, Australia, and South Africa, this case was a part of a worldwide legal experiment. Although the AI-inventorship claim was denied by many governments, South Africa was the first to grant a valid patent, with DABUS as the inventor. This raised more general concerns about the direction of AI-driven innovation and patent law going forward.

The case raised a basic question. If a machine produces anything new, who, if anyone, receives legal acknowledgement for the invention? This is similar to an AI-generated painting that lacks an artist’s signature. Due to statutory interpretation, UK courts adhered to traditional definitions, which paved the way for continuing discussions on AI’s place in intellectual property.

Legal Issues and Arguments

The main legal issue in Thaler v. Comptroller-General of Patents, Designs and Trade Marks was whether AI systems may be considered as an “inventor” under the UK Patents Act 1977. This case brought up important questions about who owns inventions created by artificial intelligence, what constitutes an inventor, and if laws should change to reflect these developments.

The Definition of “Inventor”

What is an “inventor”? Section 7(3) of the Patents Act 1977 defines an inventor as “the actual deviser of the invention”. Traditionally, the term “deviser” has been used to describe a natural person who uses their cerebral abilities to produce an invention. However, Dr. Stephen Thaler contested this interpretation, claiming that AI systems are not expressly prohibited from becoming creators by law. He argued that patent law should acknowledge the realities of contemporary technology breakthroughs and permit AI systems to be listed as inventors, as DABUS had produced innovative and practical inventions on its own.

Furthermore, Thaler contended that denying AI inventorship would compromise patent law’s incentive frameworks. By giving innovators exclusive rights, patent protection aims to promote innovation. Companies and researchers may be deterred from utilising AI in research and development if AI-generated inventions are ineligible for patent protection, which would ultimately slow down technological advancement.

The Question of Ownership

Since he owns DABUS, Thaler contended that he should be the owner of the patents for DABUS’s inventions. He asserted that DABUS should be allowed patent rights because it developed the inventions comparable to an employee, drawing a comparison between employer-employee patent rights. He also invoked the law of accession, comparing it to a cow giving birth to a calf, which asserts that the owner of an original thing owns any new product it produces. According to this reasoning, he asserted that DABUS should be entitled to their legal rights as he independently created novel concepts.

Broader Implications and Global Divergence

Concerns concerning the inconsistent treatment of AI-generated inventions in international patent law were also brought up by the case. Thaler cited South Africa as the first nation to grant a patent, with DABUS as the inventor, even though the UK rejected his claims. Furthermore, Australia reversed its verdict on appeal after first ruling in favour of AI inventorship. According to Thaler, these global variations showed that patent law is changing and that countries like the UK should update their legislation to make AI-generated ideas more clear.

Legal Certainty and Policy Concerns

The case raised wider policy issues about AI’s place in IP law in addition to the particular legal restrictions of the Patents Act. According to critics, allowing AI to be listed as an inventor may lead towards legal ambiguity surrounding patent ownership, enforcement, and accountability. Courts and patent offices would have to deal with challenging issues about accountability, inventorship disputes, and the transfer of patent rights if AI systems could hold inventorship status.

Furthermore, Thaler’s opponents stressed that the law needs to combine legal stability with technological advancement. The current patent system operates on the premise that only human inventors are capable of generating ideas and possessing legal rights, even if artificial intelligence may play an increasingly and larger role in invention. They maintained that legislative change, not judicial interpretations, should oversee any extension of patent law to allow AI inventorship.

Conclusion on Legal Issues

Beyond the technical vocabulary of patent law, the legal issues presented in Thaler v. Comptroller-General of Patents represent the more general difficulty of incorporating AI into established legal frameworks. The case finally brought to light larger policy concerns about legal certainty, ownership rights, and the changing role of AI in invention, despite Thaler’s attempts to broaden the definition of an inventor and assert ownership over patents produced by AI. Jurisdictions worldwide are under growing pressure to define how intellectual property law should change to accommodate technological advancements, and the discussion about AI inventorship is still going strong.

Court’s Decision and Reasoning

In the end, Thaler’s appeal was denied by the UK Supreme Court, which upheld the rulings of the Court of Appeal, the High Court, and the UK Intellectual Property Office (UKIPO). The Court decided that DABUS being an AI system, cannot be recognised as an inventor under the Patents Act 1977, which requires inventors to be natural persons. Legal precedent, the concept of legal certainty, and legislative interpretation served as the foundation for the Supreme Court’s reasoning.

Strict Interpretation of the Patents Act 1977

Section 7 of the Patents Act 1977 served as the foundation for the Supreme Court’s reasoning, as it specifies the eligibility to apply for and obtain a patent. The Court noted that the definition of an inventor in Section 7(3) is “the actual deviser of the invention,” which has historically been interpreted to mean natural persons.

The Court pointed out that the idea of an inventor has traditionally been associated with someone who comes up with the idea on their own. DABUS is unable to satisfy this statutory condition since, it lacks both legal personality and the capacity to use mental faculties. The Patents Act’s references to “persons” only pertains to legal or natural persons, not artificial intelligence, the Court further noted.

Inability of AI to Hold and Transfer Patent Rights

If AI cannot be an inventor, then who owns the invention? The Court decided that just because someone owns a machine doesn’t mean they have the right to a patent on the products of that machine.

Thaler’s argument that he should be able to retain the rights to the patent as the owner of DABUS was rejected by the ruling. It discovered no legal foundation for such a “derived right” from AI-generated inventions in the Patents Act of 1977 or case law that was already in existence. There is no equivalent mechanism for AI, in contrast to employer-employee interactions, where patent rights can be transferred via statute or contract.

Thaler’s reliance on the accession doctrine which applies to material property but excludes intellectual property, was also dismissed by the Court. The justices clarified that an AI system’s inventorship rights cannot be automatically transferred to its owner under UK law.

Legal Certainty and the Role of Parliament 

By declaring that the acceptance of AI as an inventor would result in significant legal ambiguity concerning inventorship, patent ownership, and enforcement, the Supreme Court underlined the necessity of legal certainty in patent law. It stated that any significant change, such permitting AI to be acknowledged as an inventor, must be made through legislative reform rather than judicial interpretation because the UK patent system is a structured, rules-based system.

The ruling clarified that the Court could not create a new category of inventorship without a legislative amendment from the Parliament and was constrained by the current statutory framework. The justices reaffirmed that their job was to enforce the law as it is, not to make assumptions about how patent law should change considering advances in artificial intelligence.

Final Ruling

The UK Supreme Court finally decided that DABUS could not be recognised as an inventor under the Patents Act 1977, because an invention must be made by a natural person. AI does not meet the legal definition of an inventor since it is incapable of exercising mental capacities or having legal personality. Additionally, the Court determined that since UK patent law does not automatically transmit inventorship rights from an AI system to its owner, Dr. Thaler, the owner of DABUS, had no right to receive patents for discoveries produced by the AI.

The Court stressed in its ruling that legislative reform, not judicial interpretation, is the only way to make any significant changes to patent law pertaining to AI inventorship. Allowing AI inventors would create legal ambiguity over inventorship, ownership, and patent enforcement under the established system of laws governing patents. The ruling preserved legal certainty, but it also raised more general concerns about how intellectual property law ought to change in light of AI-driven innovation. The Court concluded that it was its responsibility to apply the law as it stands and not to make predictions about how patent law would develop in the future.

Critical Analysis

The case’s decision upholds a human-centric approach to inventorship while posing serious concerns regarding the suitability of current patent law for dealing with innovations produced by artificial intelligence. The Court’s formalistic approach ignores the wider policy ramifications of barring AI as an inventor in a fast changing technical world, even though its rationale is consistent with a strict interpretation of the Patents Act 1977.

The decision’s strict adherence to statutory interpretation without conducting a purposeful study of whether the conventional notion of inventorship is still appropriate is one of its fundamental limitations. The Court tacitly linked inventorship to human creativity and intelligence by relying on the concept of an inventor as the “actual deviser of an invention.” However, this humanistic concept of inventorship may no longer accurately reflect technical reality as AI systems like DABUS become more and more capable of producing original, nonobvious, and useful concepts. Instead of completely banning AI-generated ideas from the patent system, the ruling ignores the issue of whether patent law should change to protect them.

From a policy standpoint, the decision discourages investment in AI-driven research by making AI-generated discoveries unpatentable, potentially impeding innovation. Contrary to the core goal of patent law, businesses may turn to trade secrecy if patent protection is denied for AIgenerated ideas, which would restrict the spread of technology. Furthermore, the fact that Australia first recognised AI patents while South Africa granted them indicates that a fragmented international patent system may develop, which would raise questions about crossborder intellectual property enforcement.

The Court was right when it said that Parliament should decide on any enlargement of inventorship, but the decision also makes clear how urgently legislative reform is needed. Creating other IP protections for works produced by AI, adding AI-assisted inventorship categories, or broadening the concept of inventorship to include AI are some possible remedies. The legislation must change in tandem with AI’s ongoing influence on innovation, or else it runs the risk of impeding rather than advancing technology.

Conclusion

While maintaining legal consistency, the decision highlights the rigidity of a human-centric system in an era of rapidly advancing AI, by reaffirming that AI cannot be an inventor under UK patent law. Although it preserves current legislation, it raises important issues regarding how AI-generated inventions may be handled by future legal systems.

This issue raises a broader policy conundrum that goes beyond law: should patent systems acknowledge AI advances, and if so, under what circumstances? A larger conflict between rigid legal definitions and developing technology is reflected in the unwillingness to address this issue. Exclusionary restrictions may become unsustainable as AI develops, necessitating a reevaluation of the law.

In the end, the case marks a significant turning point in the worldwide discourse surrounding AI and intellectual property. The argument over AI-generated patents is still ongoing, and how different countries respond in the next years will determine whether patent systems encourage or impede innovation. And so, as the law clings to the past, the future races ahead. Unburdened, unclaimed, and perhaps, for the first time in history, creating not for man, but in spite of him.

References (S):

1. Thaler v Comptroller General of Patents Trade Marks and Designs [2021] EWCA Civ 1374
2. Thaler v Comptroller-General of Patents, Designs and Trade Marks [2023] UKSC 49
3. The Patents Act 1977
4. Ed Conlon, ‘DABUS: South Africa issues first-ever patent with AI inventor’ (Managing IP, 29 July 2021) <https://www.managingip.com/article/2a5bqo2drurt0bxl7ab1q/dabus–southafrica–issues–first–ever–patent–with–ai–inventor > accessed 14 March 2025

6. David Webber, ‘High Court powers down DABUS patent prospects in Australia’ (Davies Collison Cave, 15 November 2022) <https://dcc.com/news–andinsights/high–court–powers–down–dabus–patent–prospects–in–australia/> accessed 14 March 2025
7. Thomas W Merrill, ‘ACCESSION AND ORIGINAL OWNERSHIP’ [2009] 1(2) Journal of Legal Analysis